Intel CEO says Alder Lake CPUs will have three major ‘AMD Zen-like’ architectural announcements, ARC Alchemist Xe-HPG GPUs to put pressure on NVIDIA

Under one interview with financial analyst Pierre Ferragu, Intel CEO and CFO spoke about their company’s strategy in x86 consumer, client and data center segments. The directors also detailed their implementation plans with upcoming consumer and cloud products, including Alder Lake, Diamond Rapids and Xe-HPG ARC GPUs.

Intel CEO says Alder Lake will be their AMD Zen Moment, three Zen-like architectural updates coming to consumers

Intel CEO Pat Gelsinger was asked when his company will have a Zen moment, to which he replied that Alder Lake brings three Zen-like innovations as described during its Architecture Day 2021. The three key innovations will include the very heterogeneous architecture that will include large and small nuclei. CEO states that AMD has nothing similar at their end, and with the heterogeneous approach, Alder Lake CPUs will boast high-performance cores along with an energy-efficient version of the large cores, leading to increased efficiency and higher multi-threading output.

Intel Iris Xe graphics show 24% performance using sampler feedback enabled in 3DMark Sampler feature test

Intel Alder Lake CPUs are going to be the first major x86 lineup with a hybrid and heterogeneous chip architecture.

But that’s not all, Intel is also making the next generation of architectures back in their labs, which according to the CEO will be ‘a pretty dramatic step forward and far beyond all that was talked about yet’. Pat could refer to Meteor Lake or future Cove cores that were recently leaked & is expected to bring a big push to IPC and architectural improvements, but I guess we’ll have to wait until the next IDM / architecture day for more details on them.

Q- Will there be a Zen moment for Intel?

ONE- We have a number of things going on here[…] We roll out the heterogeneous architecture that is part of Alder Lake, where we have large and small cores, you know AMD only has one. We want a higher performance and a more energy efficient version of the kernel, pretty convincing, we want to lay out our big vector enhancement stuff, we have our GPU architecture where we will start to be able to really put pressure on NVIDIA for the first time forever.

And we set up our IPU architecture when it comes to our smart NICs and smart networking materials so three major architectural announcements this week that we think are pretty zen-like.

And believe me, we have a few that are still boiling back in the labs that we’re going to look forward to talking to people about, which we think is a pretty dramatic step forward and far beyond all that was said. about yet and some that may not be talked about for a few years but innovation, you know the nerd is back.

Intel will put pressure on NVIDIA for the first time with its Xe-HPG ARC GPUs

Intel’s CEO mentioned not only AMD but also NVIDIA. Pat says that with their upcoming graphics (Xe-HPG) architecture to ARC GPUs, they will put pressure on NVIDIA’s GPU lineup for the first time ever! The company has so far only landed its Xe GPU architecture on mobile stuff, but with Xe-HPG, the company will take the fight to all segments, including desktop and discrete mobile graphics. But this is the first time the company has stated that it will compete against NVIDIA, which is one undisputed king in the discrete and AIB GPU segment.

Intel’s Xe-HPG ARC GPUs are said to offer competitive performance and pricing over NVIDIA and AMD’s discrete GPUs.

Intel looks to regain sustained leadership in cloud segments through better performance, mentions Diamond Rapids as future Xeon CPU family

Pat also talked about how they aim to regain trust in the data center and cloud segment. According to the CEO, Intel certainly acknowledges the AMD threat, and this has been the case since Brian Krzanich called AMD ‘Formidable competition‘and stated that it is their task not to let AMD capture a double-digit market share of 15-20%. Within a few years, AMD went from a market share of 0% in 2017 to over a 10% market share in 2021 with its Zen-powered EPYC series.

Intel’s CEO says they were in a period where they were behind AMD for a few years, and that was due to poor execution of their server / cloud strategy. However, they will do better, and with products such as Ice Lake and the upcoming Sapphire Rapids / Diamond Rapids Xeon chips, the company expects a period in which they will be very competitive, and in 3-4 years, Intel will take sustainable leadership. within the cloud segment.

Q How do you make sure Intel stays central and the number one client in the cloud.

ONE-There are probably three different aspects Pierre needs to talk about, you know there is always AMD viewing and AI and alternative architecture viewing, and then how we deepen the partnership with the guys of the cloud. So we seem to think about it in all three dimensions.

We’re just saying about the AMD threat, you know we need to make better products, period!

So we want to build better products and they had a period we did not perform well. We will perform, perform, perform!

and with that, we see that we are entering a period where we are going to be very competitive. So we were in a period where we were behind. Now with Ice Lake, Sapphire Rapids, Diamond Rapids, we are in a period where we are going to be very competitive and when we reach 3 -4 years from now, we see that we will be in sustainable management again.

So we are entering a period where we are going to be much more competitive and with products that you know are much more competitive, more pricing dynamics, better market share potential, better TCO values ​​for cloud guys, so we just have to compete .

Intel’s next-generation Xeon CPUs, such as the Sapphire Rapids-SP, are said to be extremely competitive with AMD’s EPYC lineup.

Intel’s Sapphire Rapids Xeon CPUs will compete with AMD’s EPYC Genoa family. The new EPYC series seems to be a huge step up compared to existing EPYC Milan chips. Intels Diamond Rapids SP The Xeon family may be the first lineup to offer leadership in the EPYC series, but is expected to be launched in 2025, which is in line with the CEO’s 3-4 year timeline.

I’m going to have the bulk of it [my product line] built internally and also be able to reap these margins. We are going to make selective use of foundries. […] If I have to use a foundry to have a unique product or a halo product and certain portions, absolutely.

I really freed up my engineering teams to say that you are going to deliver the best product. Period. You are unlimited in choosing the technologies to always deliver the best products you have in the category.

Intel also wants to take advantage of the use of both internal and external fabs with its IDM 2.0 strategy as detailed during the process and the day of manufacture. The CEO says he wants his design team to leverage the best foundries to build the best products. Overall, Intel’s new strategy under Pat Gelsinger, CEO, seems to be able to work if handled and executed well. The company faces a lot of competition from chipmakers, and its manufacturers are also undressing to deliver the next generation of nodes on time, but with this new strategy, Intel was once again able to lead the industry forward in terms of innovation!

Give a Comment